God Eater Resurrection

[Kuuyo]

This might be a shot in the dark, but I've found out it doesn't extract all the sound files I need..
To give some more information: I am using the .rdp of the Japanese VITA version of God Eater Resurrection
The files that are missing are 4 sets of .nus3bank files which follow the naming of demo101_1_pc#g (# being a number from 1 to 21, g being f or m), demo103_2_pc#g, demo108_1_pc#g, demo231_1_pc#g
As example: demo101_1_pc4f.nus3bank
There's 21m and 20f files of each of the sets.

Something else which is missing are files labeled in the format "cmn_####_Character", for example: cmn_0890_Alisa.at9
There's 956 of these, and I get only 8 of them with the script

Wondering if there's any chance for an update of the script ?

[aluigi]

What's the final coverage statistics you get at the end of the extraction?
I mean those with "coverage" and so on.

It should be visible there if there is some data that has not been parsed by the script.

Ah, how much big is the archive you are trying to extract? (in case I will need to analyze it)

[Kuuyo]

Here's the coverage statistics, I guess that also gives the files size ? (2.024.407.040 bytes)

[aluigi]

Are you using the god_eater_2.bms script?
The comment at the beginning of the script tells that the script is "Experimental, it's a scanner", indeed the code just tries to scan the whole archive searching the constant "magic" of the compressed files and decompresses them on the disk.
It may be possible that the files you mentioned aren't compressed and therefore are missed by the scanner.

There is also another thing, the extractor of Pres archives checks if "OFFSET & 0x80000000" since those not matching that "filter" were invalid. It's possible that those files are not extracted for such reason and can't be extracted anyway.

It's even possible that the files are extracted with a different name because this Pres format uses a weird way for building the filenames, I doubt but it's just another hypothesis.

Anyway the script is quite complex and supports QPCK, Pres and rdp.

Long story short, it's quite a mess.

[Kuuyo]

Yeah, that's the script I'm using.
Yep, it is quite the mess, I've been slamming my head against the .rdp for the past week trying to figure out how to get those files..
As those are the last files I need to 100% my JP Voice patch for God Eater: Resurrection on PC.

I did find a Pres which had some indexes for the demo files I'm missing but

The offset just doesn't make sense, points to the middle of something else and as you can see the offset for the next file isn't even large enough to account for the previous file..
The cmn_ series of files also have an offset starting with 0x4 and I haven't been able to make sense of it yet.

I know the files that have an offset starting with 0x8 are those that are already in the data folders outside the .rdp, so those are not in the .rdp itself.

An improvement I can see for the script already is for it to recognize .nus3bank files instead of dumping them as .nus with a bunch of garbage attached to it (like some of the .nus files have several blz4 in them too). And some nameless .wav files are actually .at9 files.

[aluigi]

The only thing I can do is analyzing that 2Gb archive, not my deepest desire but at least the filenames you provided are a good starting point for the analysis
Is it ok for you to upload it somewhere and provide the link?

[Kuuyo]

Yeah, I really don't wish upon anyone else to trudge through this file, but I'm at a point where I've stared at it for so long I should see if someone else might find the missing link..
I've pretty much unconsciously started scrolling through the .rdp every night before going to sleep for some days now, haha
This is pretty much my first real personal project, so I'm a little overwhelmed

Here's the file:
https://drive.google.com/open?id=131e76 ... M-6ljllbYj

The Pres with the information about the demo_ files is at address 1052921856/0x3EC25000
Here's a log file I generated for that Pres if it helps any: http://puu.sh/BGMWp/f7c82cf95d.txt

And as for the cmn_ files, they're scattered around it seems, and I mostly spent time on figuring out how to get the demo_ files, but when the 0x4 offset mystery is deciphered those should be solved too theoretically (hopefully)

[aluigi]

That "cmn_0890" string is contained in one of the extracted files (0001aa70.dat in the new script) that doesn't have a format altough it looks like a part of Pres without Pres information.
The only thing I did was a small update of the script that is now able to extract few more files but the coverage is the same as before.
Currently that's the max I can do.

[Kuuyo]

Thank you so much for wanting to look into it!
I actually just found a new clue too, the 00002e30.dat seems to contain some of the demo103_2_pc#g nus3bank files!
There's some blz4 in there too, do those ever get unpacked? What exactly does the name of the files outside folders mean? Is there a way to backtrace where they came from?
Sorry for the waterfall of questions, but I'm kind of hyped up now and hope this might bring us a step closer to unraveling the mystery!

[Kuuyo]

It actually seems all of the 00002e## files might be all of the missing .nus3bank demo files..
Wow, I feel like doing a little victory dance for finally having some progress again

[Kuuyo]

It seems that except for 00002e30 and 00002e2f, the sizes of the .dat files are what the .nus3bank files are supposed to be, but the nus3bank file itself doesn't start before address 0x00010000/65536, so 65536 bytes of the nus3bank is (in the next .dat file in the sequence)

EDIT: They are not, the end of the file is at the start of the file

[Kuuyo]

Files 00002e06 to 00002e2f contained the full demo101_1_pc#g.nus3bank set of files!
As for 00002e30, it contained some of demo103_2_pc#g and all demo231_1pc#f (yes, only the f, no m) files
I wonder if the remaining ones are in the blz4s in between them

[Kuuyo]

(I've tried fiddling with the script to only get the BLZ_EXTRACT, but I haven't been successful..
Would there be a way to have it scan in the .dat files for blz4 or even just have a script to unpack blz4 when they're standalone?
Since I think I was able to get those out too while I cut out the .nus3bank files
http://puu.sh/BHapT/50189f5867.rar)

EDIT: Disregard this, I found your Offzip tool, thanks!

[Kuuyo]

The blz4 files in there did indeed hold the still missing demo files, I'm still processing them which I'll only be able to complete tomorrow so I'm not sure if I got them all
I have a stinking suspicion 00002e2f.dat is demo103_2_pc1f.nus3bank end which would be extracted first from a blz4, as the first blz4 in 00002e30.dat holds demo103_pc1m.nus3bank, but I wonder where the rest of demo103_2_pc1f.nus3bank went then..

[Kuuyo]

There was one part of demo103_2_pc1f.nus3bank that I indeed couldn't find, but I found out the location for it in the .rdp, so I've obtained all demo files !
Now just heap of cmn_ files left, oh boy..

[aluigi]

You are doing a lot of work
So do you still need support on this archive?

[Kuuyo]

Thanks for the offer, it's really appreciated !
I feel like only having had contact with you inspired me to break through the slump I was having, LOL
I just finished analyzing the structure of how and where the cmn_ files are stored, best case scenario they're all in that block I have my eyes set on now
Have the steps a program would need to extract them written down on a paper too, so I just need to write that and I might be done !

Thanks a lot though aluigi, the script has helped me a lot with extracting and understanding the archive, I most likely would never have gotten this far without it!